Heart failure is the physiological state in which cardiac output is insufficient to meet the needs of the body or to do so only at a higher filing pressure. There are many underlying causes of HF, including myocardial infarction, coronary artery disease, valvular disease, hypertension, and myocarditis. Chronic heart failure is associated with neurohormonal activation and alterations in autonomic control. Although these compensatory neurohormonal mechanisms provide valuable support for the heart under normal physiological circumstances, they also play a fundamental role in the development and subsequent progression of HF.
For example, one of the body's main compensatory mechanisms for reduced blood flow in HF is to increase the amount of salt and water retained by the kidneys. Retaining salt and water, instead of excreting it via urine, increases the volume of blood in the bloodstream and helps to maintain blood pressure. However, the larger volumes of blood also cause the heart muscle, particularly the ventricles, to become enlarged. As the heart chambers become enlarged, the wall thickness decreases and the heart's contractions weaken, causing a downward spiral in cardiac function. Another compensatory mechanism is vasoconstriction of the arterial system, which raises the blood pressure to help maintain adequate perfusion, thus increasing the load that the heart must pump against.
In low ejection fraction (EF) heart failure, high pressures in the heart result from the body's attempt to maintain the high pressures needed for adequate peripheral perfusion. However, as the heart weakens as a result of such high pressures, the disorder becomes exacerbated. Pressure in the left atrium may exceed 25 mmHg, at which stage, fluids from the blood flowing through the pulmonary circulatory system transudate or flow out of the pulmonary capillaries into the pulmonary interstitial spaces and into the alveoli, causing lung congestion and if untreated the syndrome of acute pulmonary edema and death.
Table 1 lists typical ranges of right atrial pressure (RAP), right ventricular pressure (RVP), left atrial pressure (LAP), left ventricular pressure (LVP), cardiac output (CO), and stroke volume (SV) for a normal heart and for a heart suffering from HF. In a normal heart beating at around 70 beats/minute, the stroke volume needed to maintain normal cardiac output is about 60 to 100 milliliters. When the preload, after-load, and contractility of the heart are normal, the pressures required to achieve normal cardiac output are listed in Table 1. In a heart suffering from HF, the hemodynamic parameters change (as shown in Table 1) to maintain peripheral perfusion.
TABLE 1ParameterNormal RangeHF RangeRAP (mmHg)2-6 6-20RVSP (mmHg)15-2520-80LAP (mmHg) 6-1215-50LVEDP (mmHg) 6-1215-50CO (liters/minute)4-82-6SV (milliliters/beat) 60-10030-80
HF is generally classified as either systolic heart failure (SHF) or diastolic heart failure (DHF). In SHF, the pumping action of the heart is reduced or weakened. A common clinical measurement is the ejection fraction, which is a function of the blood ejected out of the left ventricle (stroke volume) divided by the maximum volume in the left ventricle at the end of diastole or relaxation phase. A normal ejection fraction is greater than 50%. Systolic heart failure generally causes a decreased ejection fraction of less than 40%. Such patients have heart failure with reduced ejection fraction (HFrEF). A patient with HFrEF may usually have a larger left ventricle because of a phenomenon called “cardiac remodeling” that occurs secondarily to the higher ventricular pressures.
In DHF, the heart generally contracts normally, with a normal ejection fraction, but is stiffer, or less compliant, than a healthy heart would be when relaxing and filling with blood. Such patients are said to have heart failure with preserved ejection fraction (HFpEF). This stiffness may impede blood from filling the heart and produce backup into the lungs, which may result in pulmonary venous hypertension and lung edema. HFpEF is more common in patients older than 75 years, especially in women with high blood pressure.
Both variants of HF have been treated using pharmacological approaches, which typically involve the use of vasodilators for reducing the workload of the heart by reducing systemic vascular resistance, as well as diuretics, which inhibit fluid accumulation and edema formation, and reduce cardiac filling pressure. No pharmacological therapies have been shown to improve morbidity or mortality in HFpEF whereas several classes of drugs have made an important impact on the management of patients with HFrEF, including renin-angiotensin antagonists, beta blockers, and mineralocorticoid antagonists. Nonetheless, in general, HF remains a progressive disease and most patients have deteriorating cardiac function and symptoms over time. In the U.S., there are over 1 million hospitalizations annually for acutely worsening HF and mortality is higher than for most forms of cancer.
In more severe cases of HFrEF, assist devices such as mechanical pumps are used to reduce the load on the heart by performing all or part of the pumping function normally done by the heart. Chronic left ventricular assist devices (LVAD), and cardiac transplantation, often are used as measures of last resort. However, such assist devices typically are intended to improve the pumping capacity of the heart, to increase cardiac output to levels compatible with normal life, and to sustain the patient until a donor heart for transplantation becomes available. Such mechanical devices enable propulsion of significant volumes of blood (liters/min), but are limited by a need for a power supply, relatively large pumps, and pose a risk of hemolysis, thrombus formation, and infection. Temporary assist devices, intra-aortic balloons, and pacing devices have also been used.
Various devices have been developed using stents to modify blood pressure and flow within a given vessel, or between chambers of the heart. For example, U.S. Pat. No. 6,120,534 to Ruiz is directed to an endoluminal stent for regulating the flow of fluids through a body vessel or organ, for example, for regulating blood flow through the pulmonary artery to treat congenital heart defects. The stent may include an expandable mesh having lobed or conical portions joined by a constricted region, which limits flow through the stent. The mesh may comprise longitudinal struts connected by transverse sinusoidal or serpentine connecting members. Ruiz is silent on the treatment of HF or the reduction of left atrial pressure.
U.S. Pat. No. 6,468,303 to Amplatz et al. describes a collapsible medical device and associated method for shunting selected organs and vessels. Amplatz describes that the device may be suitable to shunt a septal defect of a patient's heart, for example, by creating a shunt in the atrial septum of a neonate with hypoplastic left heart syndrome (HLHS). That patent also describes that increasing mixing of pulmonary and systemic venous blood improves oxygen saturation, and that the shunt may later be closed with an occluding device. Amplatz is silent on the treatment of HF or the reduction of left atrial pressure, as well as on means for regulating the rate of blood flow through the device.
Implantable interatrial shunt devices have been successfully used in patients with severe symptomatic heart failure. By diverting or shunting blood from the left atrium (LA) to the right atrium (RA), the pressure in the left atrium is lowered or prevented from elevating as high as it would otherwise (left atrial decompression). Such an accomplishment would be expected to prevent, relieve, or limit the symptoms, signs, and syndromes associated of pulmonary congestion. These include severe shortness of breath, pulmonary edema, hypoxia, the need for acute hospitalization, mechanical ventilation, and death.
Shunt flow is generally governed by the pressure gradient between the atria and the fluid mechanical properties of the shunt device. The latter are typically affected by the shunt's geometry and material composition. For example, the general flow properties of similar shunt designs have been shown to be related to the mean interatrial pressure gradient and the effective orifice diameter.
Percutaneous implantation of interatrial shunts generally requires transseptal catheterization immediately preceding shunt device insertion. The transseptal catheterization system is placed from an entrance site in the femoral vein, across the interatrial septum in the region of fossa ovalis (FO), which is the central and thinnest region of the interatrial septum. The FO in adults is typically 15-20 mm in its major axis dimension and ≤3 mm in thickness, but in certain circumstances may be up to 10 mm thick. LA chamber access may be achieved using a host of different techniques familiar to those skilled in the art, including but not limited to: needle puncture, stylet puncture, screw needle puncture, and radiofrequency ablation. The passageway between the two atria is dilated to facilitate passage of a shunt device having a desired orifice size. Dilation generally is accomplished by advancing a tapered sheath/dilator catheter system or inflation of an angioplasty type balloon across the FO. This is the same general location where a congenital secundum atrial septal defect (ASD) would be located.
U.S. Patent Publication No. 2005/0165344 to Dobak, III describes apparatus for treating heart failure that includes a tubular conduit having a emboli filter or valve, the device configured to be positioned in an opening in the atrial septum of the heart to allow flow from the left atrium into the right atrium. Dobak discloses that shunting of blood may reduce left atrial pressures, thereby preventing pulmonary edema and progressive left ventricular dysfunction, and reducing LVEDP. Dobak describes that the device may include deployable retention struts, such as metallic arms that exert a slight force on the atrial septum on both sides and pinch or clamp the device to the septum.
Two types of percutaneously implantable shunts have been described in the medical and patent literature. In short-term, small-size clinical trials, both types have been shown to be associated with improvements in symptoms, quality of life measurements, and exercise capacity. Both shunts also have observed and theoretical drawbacks, which may limit their effectiveness and use.
The first type of shunt is henceforth referred to as an orifice-plate mesh shunt. Orifice-plate mesh shunts comprise a metallic mesh that wraps around both sides of the septum with a hole in the center and anatomically mimics the location and geometrical characteristics of a small congenital secundum ASD. The shunt geometry generally resembles a thin plate with a hole in it. In most embodiments, the “plate” comprises both mesh material and atrial septal tissue encased by the mesh. One example of such devices, designed by Corvia Medical, Inc., Tewksbury Mass., consists of a self-expanding nitinol mesh that forms a pair of disc-like flanges with an open orifice in the center. The maximal diameter of the discs is 19.4 mm and the orifice diameter is 8 mm. Each disc flange has multiple truss-like legs that deploy into a preset configuration that wraps around the LA and RA sides of the interatrial septum and applies a clamping force to the tissue.
Another example of such a mesh type device, developed by Occlutech International AB, Helsingborg, Sweden, resembles a dual-disc occluder used for closing congenital secundum ASDs, which additionally includes a short open barrel orifice in the center that connects the two discs.
A major benefit of the foregoing orifice-plate mesh shunts over other shunt designs is simplicity of manufacture. Although relatively simple in theory and construction, orifice-plate mesh type shunts have several important drawbacks that are expected to reduce their overall potential for clinical safety and effectiveness.
A first drawback of orifice-plate devices is the susceptibility to narrow or close during the post-implantation healing period. For example, neoendocardial tissue ingrowth, referred to as pannus, grows from the underlining tissue to cover the mesh and narrow or partially occlude the shunt orifice. During the period following implantation, local trauma caused by crossing and dilating the FO, plus the chronic effects of continuous pressure applied by the mesh material on the septal tissue, provoke a localized healing response. This response entails activation of an inflammatory process, attracting lymphocytes and macrophages to the area of tissue injury. These inflammatory cells in turn release a variety of cytokines that signal fibroblasts and smooth-muscle cells from the wound margins to dedifferentiate, migrate, proliferate and encapsulate affected portions of the implanted device. The fibroblasts and smooth muscle cells then secrete extracellular matrix material composed of collagen and proteoglycans, which extracellular matrix forms the bulk of the pannus. The duration of this healing phase in humans is typically up to 6-9 months, but may be longer if there is a chronic source for tissue injury such as device compression or erosion of adjacent tissue. Eventually this pannus is covered with neoendothelial cells, causing the pannus growth to stop or stabilize. In the long term, the collagen of the pannus remodels, but generally retains its space occupying properties. Such tissue ingrowth typically spreads over the surfaces of the implant's struts, mesh, or discs, and may substantially narrow the orifice lumen or even entirely occlude the shunt. Narrowing or occlusion of the shunt prevents LA decompression and limits any positive effect for the patient.
The degree of luminal narrowing may be quite variable between patients due to differences in the severity of local injury—the more injury, the more exaggerated the pannus formation. Also, variability results from differences in host wound healing responses. For example, the amount and character of extracellular matrix may affect the duration of healing and amount of material deposited. Thus, for an orifice-plate mesh shunt, the eventual orifice lumen size will be highly variable. These processes will be familiar to one skill in the art as it is generally analogous to the type of late lumen loss that occurs in arteries when bare metal stents are used to treat atherosclerotic stenosis.
In a trial described in the publication, “A Transcatheter Intracardiac Shunt Device for Heart Failure with Preserved Ejection Fraction (REDUCE LAP-HF): A Multicentre, Open-label, Single-arm, Phase 1 Trial” by Hasenfuss, et al., 14 of 64 patients implanted with an orifice-plate mesh shunt device had no demonstrable flow across the shunt on transthoracic echocardiographic Doppler imaging at 6 months after implantation. It has not reported whether the shunts were occluded or if the imaging study was simply too technically difficult to tell for certain. Although additional interventional cardiology procedures may be undertaken to restore lost luminal patency, such procedures may pose unacceptable risks, including death and stroke from embolization of the orifice-clogging material.
A second drawback of an orifice-plate mesh shunt is the potential for paradoxical embolization. Paradoxical embolization refers to thromboembolism originating in the venous vasculature (venous thromboembolism or VTE), such that an embolus traverses right-to-left through a cardiac shunt into the systemic arterial circulation. The most severe complication of paradoxical embolization occurs when an embolus lodges in the cerebral circulation with resulting cerebral infarction (stroke). Similarly, if a paradoxical embolus enters the coronary arterial circulation, myocardial infarction (MI) may ensue. Other embolic syndromes result from embolization to the mesenteric, renal, and peripheral arteries supplying the limbs. These may cause respectively, ischemic bowel syndrome, hematuria with worsening renal function, and gangrene requiring amputation.
Most frequently, VTE in adults is the consequence of in situ thrombosis in the deep veins (deep venous thrombosis or DVT) of the lower extremities or pelvis. For the most part, clinically relevant venous emboli develop in the popliteal veins or more proximally in larger veins of the upper thigh or pelvis. In patients with DVT involving the popliteal vein, the venous diameter averaged 11.4 mm (range from 6.2 mm to 20.1 mm). Often, emboli are described as having the form of a cast of the vein's lumen with a width equal to the diameter of the vein of origin. These thrombi also tend to be elongated, corresponding to the length of the occluded venous segment.
The risk factors associated with thromboembolic disease include a variety of anatomic, physiological, rheological variables and disease states. Heart failure is a well-recognized risk factor for DVT and VTE, especially in patients with reduced left ventricular systolic function. About 3% of deaths in heart failure patients are due to VTE, usually associated with pulmonary embolism. Patients with transvenous endocardial pacing leads and an intracardiac shunt have a 3-fold increased risk of systemic thromboembolism, suggesting that paradoxical embolism is a contributing underlying cause. There is evidence that the risk of paradoxical embolism is directly related to the orifice size of naturally occurring atrial level shunts such as ASD and patent foramen ovale (PFO). The presence of an atrial septal aneurysm is an additional risk factor. For example, as described in the publication “Transcatheter Amplatzer Device Closure of Atrial Septal Defect and Patent Foramen Ovale in Patients with Presumed Paradoxical Embolism” by Khositsth, et al., in a series of 103 adult patients with paradoxical embolization, an ASD was present in 12%, whereas PFO was present in 81%. In patients with clinically significant ASD referred for closure, the incidence of paradoxical embolus has been reported to be up to 14%.
It has been asserted that in order for VTE to enter the systemic circulation, the prevailing LA to RA pressure gradient must be temporarily reduced, eliminated or reversed so that blood will either flow slowly across the shunt, cease to flow across the shunt or flow retrograde across the shunt. Echo/Doppler imaging studies often reveal some amount of shunting in both directions (bi-directional shunting) in patients with congenital ASD, even when LA to RA flow predominates. Bidirectional shunting may be best demonstrated when a subject performs a Valsalva maneuver (straining caused by exhalation against a closed glottis). Valsalva increases intrathoracic pressure, which causes the RA and LA pressures to equalize after several seconds and then for the RA pressure to transiently exceed LA pressure on exhalation. Intermittent bidirectional flow also may be observed at rest when the interatrial pressure gradient is low, or intermittently during the cardiac cycle when LA contraction is delayed compared to RA contraction (interatrial conduction delay). This is seen especially when the atria are enlarged or diseased, such as in heart failure. In this setting, interatrial electrical conduction delay results in retardation of LA contraction. Bidirectional shunting can also be seen transiently during inspiration, when venous return to the RA is increased, during coughing, with abdominal compression, during forced exhalation, or in the presence of severe tricuspid valve regurgitation. Chronically increased pulmonary arterial pressure, as seen in severe pulmonary hypertension, whether primary or secondary to chronic lung disease, recurrent pulmonary embolism, or due to chronic right ventricular volume overload, has been associated with chronic and more severe RA to LA shunting.
Additional phenomena associated with RA to LA shunting are diminished pulmonary blood flow and decreased arterial oxygen saturation due to systemic venous admixing. When these findings are also transient, they are generally well tolerated. Thus, prevention of significant or larger paradoxical emboli is the primary concern rather than preventing reverse shunting per se. As the consequences of paradoxical embolization can be catastrophic, it is desirable particularly in high-risk patients, that an implantable shunt be equipped with mechanism(s) that limit or minimize the chances of paradoxical embolization or minimize the chances of transporting large emboli.
From these data, it seems reasonable to expect that an orifice-plate mesh shunt, by virtue of its anatomic similarities with congenital secundum ASD, would have a similar risk of paradoxical embolization. It is easily understandable that a thin plate-orifice mesh type of artificial shunt might be more susceptible to paradoxical embolization than other types of shunts with longer orifice geometries, e.g., a nozzle. For any given quanta of RA volume (blood or thrombus), the statistical likelihood of traversing retrograde across the shunt and into the LA would be expected to be a complex function of the duration of pressure gradient reversal, flow patterns in the RA, shunt tunnel distance affecting the length of the flow velocity streamlines, and flow velocity and orifice or lumen size.
A third drawback of an orifice-plate mesh shunt is that percutaneous removal from the shunt body is only possible at the time of implantation. Should the shunt become a nidus for infection, develop fatigue or corrosion fractures of its metallic framework, or erode or otherwise impinge on other vital cardiac structures, it cannot be removed by percutaneous retrieval/removal techniques. This is because the shunt, with its large “footprint” on the interatrial septum, is encased in pannus tissue. Attempts at percutaneous removal may result in tearing of the septum, pericardial tamponade, and device embolization into the systemic circulation, resulting in death or the need for emergency surgery. Safe removal would require performing open heart surgery. This entails that the heart be bypassed using an extracorporeal membrane pump oxygenator (cardiopulmonary bypass), so the heart can be opened, the shunt removed, and the septum repaired. Performing such surgical procedures in patients with already established severe heart failure, including its frequently associated co-morbid conditions such as peripheral, cerebrovascular, and coronary artery disease, renal dysfunction, and diabetes, would be expected to have substantial risks for mortality or severe morbidity.
A fourth drawback of an orifice-plate mesh type of shunt is that its geometry renders it relatively inefficient in supporting high flow. For any given pressure gradient across the shunt, the geometry of an orifice plate requires a larger orifice because it has a reduced effective orifice size compared with other geometries, such as a venturi-shaped lumen, or a conical shaped nozzle. This is because with an office-plate, there are more energy losses associated with eddy currents at the edges of the plate. Orifice-plate geometries may be categorized as having a relatively low discharge coefficient, which is a dimensionless fluid-mechanical parameter that relates flow to actual orifice size. For practical purposes, the discharge coefficient is the ratio of areas of the exiting jet vena contracta, which is the narrowest portion of the jet, compared to the shunt orifice. For example, the coefficient of discharge for orifice plates placed in pipes tends to be approximately 0.6, but rarely exceeds 0.65. The discharge coefficient is affected by the orifice and chamber dimensions, the pressure gradient, and the viscosity of blood and/or the Reynolds number of the specific flow condition. This differs from the more efficient passage of flow through a classic venturi type of narrowing, where the discharge coefficient usually exceeds 0.9 and is typically in the range of 0.94 to 0.98. The result is that, in comparison with more efficient shunt lumen geometries, an orifice-plate mesh shunt requires a larger orifice diameter to accommodate the same amount of flow for any given pressure differential across the shunt.
A fifth drawback of an orifice-plate mesh shunt is that it occupies a large area or footprint on the interatrial septum. The flanges of the device that anchor the shunt typically occupy the entire area of the fossa ovalis and may overlap adjoining muscular portions of the interatrial septum. These flanges exert persistent pressure on the septum, causing injuring and stimulating an exaggerated healing response as described above. Also, the rigidity of the mesh may interfere with the normal motion of the muscular septum. The flanges additionally may impinge on adjacent cardiac structures such as the roof of the left atrium, the ostia of the pulmonary veins, and the aorta root and sinuses of Valsalva, where due to chronic rubbing contact or sandwiching compressive forces, they may erode into these vital structures. Such erosion has been associated with severe complications including cardiac tamponade and death. For example, the similarly sized Amplatzer ASD disc occlusion device described above has been occasionally associated with erosion into adjoining tissues with resulting catastrophic outcomes.
Additional issues associated with placing relatively large devices with complex three-dimensional geometries are potential difficulties in positioning the shunts accurately in the FO, obtaining sufficient tissue anchoring to prevent migration, and having devices conform to irregularities of the cardiac anatomy. For example, in a report of attempted implantation of orifice-plate mesh shunts in 66 patients in the above cited publication authored by Hasenfuss, et al., device placement was not possible in two patients. And of the 64 implanted patients, the device had to be removed and re-implanted in another 3 patients due to misplacement, migration, or embolization of the first attempted implant.
Finally, the large footprint on the atrial septum may hinder or render impossible performing other interventional procedures that require transseptal access. The large flange diameter and small mesh pore sizes generally make catheter crossing of the atrial septum possible only through the central shunt orifice itself. Transseptal procedures using small diameter catheters, such as atrial fibrillation RF ablation, may be conducted through the orifice-plate lumen only if it is not obstructed by pannus and the orifice location permits entry into all four pulmonary veins. Other structural heart disease procedures that have large diameter delivery systems and/or require crossing the FO in specific locations may encounter difficulties or simply not be possible. These procedures include left atrial appendage occlusion, mitral valve edge-to-edge (“MitraClip”) repair, and transvascular mitral valve replacement. For example, placing of a MitraClip optimally requires crossing the FO in its superior-posterior quadrant. The guiding catheter has a tip inner diameter of 7.7 mm (23 Fr). Similar transseptal access is needed to perform reconstructive mitral annuloplasty with the Cardioband device marketed by Valtech. In these cases, the only alternatives might be higher risk therapeutic approaches involving trans-left ventricular apical access or open heart surgery.
The second type of shunt is referred to as a valved unidirectional shunt. These shunts attempt to overcome some of the drawbacks of orifice-plate devices. For example, valved unidirectional shunts have embodiments containing a one-way or check-valve to limit reverse shunting and paradoxical embolization. Some of the valve configurations are designed to open when the LA-RA pressure gradient exceeds a predefined threshold. Other valve configurations close only when the RA pressure exceeds LA pressure (reversed gradient).
U.S. Pat. No. 9,034,034 to Nitzan, the entire contents of which are incorporated by reference herein, solves many of the drawbacks of plate-like orifice mesh shunts describe above. An embodiment of the Nitzan-type shunt comprises an hourglass or diabolo outer shape, having a small FO footprint minimizing septal injury, which is expected to minimize pannus growth and obliteration of the shunt lumen. Its one-way valve also is designed to reduce the potential for reverse shunting and paradoxical embolization. The relatively small footprint of the shunt in contact with the septum and encapsulated collapsible nitinol frame is designed to facilitate percutaneous extraction from the septum and retrieval from the body using a standard goose-neck snare and large-bore sheath, thus making the device more easily retrieved. The venturi tube-like inner lumen of the diabolo shape provides better bulk flow characteristics, permitting a smaller orifice for the same amount of flow compared to orifice plate shunts. And finally, the small footprint on the FO and the hourglass shape are designed to facilitate accurate placement and retention during implantation. This geometry also minimizes interference with normal motion of the interatrial septum, and the small footprint provides space surrounding the shunt for other potential interventional procedures that require transseptal catheterization.
One embodiment of the Nitzan design, manufactured by V-Wave, Ltd (Caesarea, Israel), designed to support unidirectional left-to-right flow, comprises a self-expanding frame constructed from a laser-cut nitinol tube. The frame includes five sinusoidal circumferential struts interconnected by six longitudinal bars. The frame is heat-set so that it has an asymmetrical hourglass shape or a diabolo shape. The shunt is deployed so that the neck (5.3 mm outer diameter) is placed across the FO and secured in place by its external surface geometry. The shunt's widest portion has a conical shape with an approximately 14.3 mm outer diameter at the LA end of the shunt, which serves as an “entry” port on the distal end of the entry funnel. The entry funnel is deployed in the left atrium, and registers the neck of the shunt to the region of the FO. A second, slightly narrower bell-shaped portion forms the exit portion of the shunt, which expands to a maximum outer diameter of 11.4 mm at the RA end of the shunt. The shunt does not require flanges, discs, or tissue anchors to secure it in place. Septal retention is achieved without applying persistent pressure, tension or rubbing contact on the tissue adjoining the device neck.
The V-Wave shunt has a single inner lumen where flow is entrained into the entry funnel in the LA and passes through the constricted neck having a 5.1 mm inner diameter, which resembles a venturi-type orifice, and then exits through a bioprosthetic valve positioned near the RA end of the shunt. The entry funnel and the central neck region are encapsulated with expanded polytetrafluoroethylene (“ePTFE”) to form a skirt or cover over the frame. The skirt is designed to facilitate laminar flow and limit pannus ingrowth during device healing. The exit bell-shaped portion contains three, glutaraldehyde-fixed, porcine pericardial leaflets sutured to the frame at the right atrial extent of the ePTFE encapsulation. The leaflets are designed to create a smooth exit channel and remain in the open position, closing only when the RA pressure exceeds LA pressure by 1-2 mmHg, thus preventing reverse right-to-left shunting.
For deployment, the V-Wave shunt is compressed in a loading tube where it is attached to a triple-latch cable delivery catheter. The loading tube is inserted into a 14F delivery sheath that has been previously placed after a transseptal catheterization from the right femoral vein across the FO. The shunt then is advanced through the sheath until the entry funnel has been deployed in the LA. The entire system is withdrawn as a unit until the LA funnel is in contact with the left side of the FO. The delivery catheter latches are unhooked from the shunt, the delivery catheter withdrawn so the right atrial side of the shunt is held only by its radial force against the delivery sheath. Then the delivery sheath is withdrawn, thereby deploying the exit bell-shaped portion of the shunt on the RA side of the FO. Device placement may be guided and confirmed by fluoroscopy and echocardiography, e.g., intracardiac echo or transesophageal echo.
Pre-clinical testing on the V-Wave shunt was performed in an established juvenile ovine (sheep) model that created an ischemic cardiomyopathy form of heart failure. The sheep were pre-treated with sequential coronary artery microembolization as described in the publication, “Chronic Heart Failure Induced by Multiple Sequential Coronary Microembolization in Sheep” by Schmitto et al. After several weeks, the sheep manifested evidence of severe left ventricular systolic dysfunction and develop elevated LV, LA, and pulmonary artery pressures. In a 12-week survival study, this V-Wave shunt was associated with significant improvements in LA pressure and left ventricular ejection fraction. All manifestations of worsening heart failure were improved and in some cases reversed with interatrial shunting. Concurrent control animals with established heart failure, but were not implanted with the V-Wave shunt, demonstrated progressive worsening of LV ejection fraction, and intracardiac/pulmonary pressure during 3-month follow-up. The physiological improvements in shunted animals were substantial even though the shunt volume was assessed to be small. The pulmonary blood flow/systemic blood flow ratio (Qp/Qs) was between 1.1 to 1.2, as measured by oximetry, which is consistent with a very small shunt. Naturally occurring ASDs, with a Qp/Qs less than 1.5, are generally left untreated as they are well tolerated for decades by the compliant right heart and pulmonary vasculature, without evidence of worsening right ventricular failure despite mild chronic volume overload. This was confirmed in the sheep model where RA and pulmonary artery pressures decreased to baseline levels with shunting, but progressively worsened in the control animals.
A total of 38 patients were implanted with the V-Wave hourglass-shaped shunt having valve leaflets in two similar feasibility studies. The baseline characteristics of the combined study populations are summarized in Table 1 below.
TABLE 1Baseline characteristics of 38 patients implanted with valved hourglass-shaped shunt deviceAge, years66 ± 9 Male gender, %92Body mass index, kg/m230 ± 6 NYHA class, medianIII (97%), IV (3%)Ischemic Cardiomyopathy, %76DM/HTN/AFIB, %68/84/53ACEi-ARB/BB/MRA/DIUR, %78/100/75/94CRT-D or ICD/CRT-D or CRT-P, %74/39NT-proBNP, pg/ml2640 ± 2301eGFR, mL · min−1 · 1.73 m−254 ± 206MWT, m282 ± 114PCWP, mmHg20 ± 6 RAP, mmHg8 ± 4PAP mean, mmHg30 ± 7 CI, L · min−1 · m−22.1 ± 0.5PVR, mmHg/L · min−12.9 ± 1.4LVEF (HFrEF, n = 30), %26 ± 7 LVEF (HFpEF, n = 8), %50 ± 9 NYHA = New York Heart Association heart failure classification;DM = diabetes mellitus;HTN = hypertension;AFIB = atrial fibrillation;ACEi-ARB = receiving angiotensin converting enzyme inhibitor or angiotensin receptor blocker;BB = receiving beta blocker;MRA = receiving mineralocorticoid antagonist;DIUR = receiving loop diuretic;CRT-D = implanted with combination cardiac resynchronization therapy pacemaker with ICD;ICD = implantable cardioverter/defibrillator;CRT-P = implanted with cardiac resynchronization therapy pacemaker without combination ICD;NT-proBNP = N-terminal pro b-type natriuretic peptide;eGFR = estimated glomerular filtration rate; 6MWT/minute walk test distance;PCWP = pulmonary capillary wedge pressure;RAP = right atrial pressure;PAP = pulmonary artery pressure;CI = cardiac index;PVR = pulmonary vascular resistance;LVEF = left ventricular ejection fraction;HFrEF = heart failure with reduced ejection fraction;HFpEF = heart failure with preserved ejection fraction.These parameters and abbreviations are well known to one skilled in the art.
All patients had New York Heart Association (NYHA) Class III or ambulatory Class IV heart failure symptoms at the time of study enrollment. Patients with either reduced or preserved left ventricular ejection fraction were included. There was a high frequency of co-morbidities known to be associated with a poorer prognosis including coronary artery disease, diabetes mellitus, atrial fibrillation, and chronic kidney dysfunction. All patients received appropriate guideline-driven medical and device therapies prior to study enrollment. Patients had evidence of elevated levels of natriuretic peptides, reduced exercise capacity, elevated intracardiac and pulmonary artery pressures, increased pulmonary vascular resistance, and reduced cardiac output. These factors have also been associated with poor outcomes. Patients were excluded if they had severe right ventricular dysfunction or severe pulmonary hypertension.
Implantation of the V-Wave shunt was successful in all 38 patients and no device replacements were performed. Shunts remained implanted in the atrial septum without dislodgements, migrations or apparent interference with normal septal motion on fluoroscopic or echocardiographic imaging. No shunts have required removal or replacement for infection or strut fracture. Follow-up imaging studies show that there are adjacent locations on the FO, that are available and amenable for performing transseptal procedures to treat other cardiac conditions, including, for example, atrial fibrillation ablation, left atrial appendage occlusion, and mitral valve repair. The valve apparatus, when functioning normally, has been shown to effectively prevent reverse (right-to-left) shunting. Echocardiographic contrast and Doppler studies during rest or Valsalva maneuver show that there is no reverse shunting in the early months after human implantation. Furthermore, no thromboembolic clinical events, including paradoxical embolization, have been observed during the first year of follow-up.
Shunt patency is defined as LA to RA flow through the shunt as observed during transesophageal echo/Doppler study. At 3-months after implantation of the V-Wave shunts, patency was confirmed in all patients. The pulmonary to systemic flow ratio (Qp/Qs), as measured by echocardiography, increased from 1.04±0.22 at baseline to 1.18±0.16 shortly after implantation (p<0.03). In the subgroup of 30 patients with HFrEF presented by Dr. William Abraham, MD at TCT 2016 in Washington D.C., there were statistically significant (p<0.05) improvements in clinician-assessed symptoms, patient assessed quality-of-life scores, and exercise capacity as measured by a 6-minute hall walk testing at 3, 6, and 12 months following implantation There was no deterioration in natriuretic hormone levels, echocardiographic, or hemodynamic parameters. Most importantly, the annualized (Poisson) heart failure hospitalization rate with shunting (0.17 heart failure hospitalization per patient year), was substantially reduced in comparison to a well matched historical control groups (CHAMPION trial Control and Treatment groups, 0.90 and 0.67 heart failure hospitalization per patient year, respectively). These data provide adequate proof-of-concept that interatrial shunting is of benefit to patients with severe symptomatic heart failure. Moreover, these data strongly support moving forward with larger-scale clinical trials including randomized clinical trials.
Notwithstanding the initial success observed in the foregoing trial, device occlusion, e.g., shunts having undetectable LA to RA flow, was observed in some valved interatrial shunt devices after long-term implantation, e.g., by 1 year. Further, shunts may develop bidirectional shunting that was not present early on. Bidirectional shunting is indicative of an incompetent valve, e.g., a valve where one or more leaflets do not fully coapt during closure, resulting in an open channel for reversed flow, and depending on the severity of the incompetence, may create a potential path for paradoxical embolus to traverse from the RA to LA.
To assess the effective orifice size of valved shunts over time, transesophageal echo/Doppler measurements of the diameter of the vena contracta, measured on the left-to-right flow jets through the shunt, were found to be consistent with progressive shunt narrowing. The vena contracta diameter monotonically decreased with time after implantation from 4.0±1.1 mm just after implantation, to 3.6±1.0 mm at 3 months, and 2.7±1.4 mm at 6-12 months (p<0.01). This equates, on average, to shunts losing more than half of their orifice area by 12 months. Moreover, some of the left-to-right jets appeared to be exiting the shunt at an angle substantially different from the long axis of the shunt body. This skewing of the jet is consistent with material inside the shunt such as a valve leaflet with impaired mobility, which diverts the direction of the jet. This observation gives rise to concern about a decrease in the clinical effectiveness of the shunts over time
Clinical effectiveness also may be measured by the rate of hospitalization for worsening heart failure. In the 38 patients, during the first 6 months after implantation of the V-Wave shunt, the hospitalization rate was 0.16 per patient year, which increased to 0.40 per patient year between months 6-12. These data suggest there may be a loss of shunting benefit consistent with the time course associated with shunt narrowing or occlusion.
There are several possible mechanisms working alone or in combination that could explain these observations.
The least likely cause of shunt occlusion is collapse of the shunt due to external forces applied by the septum. For example, it is possible that contraction of pannus tissue formed during the later stages of healing (remodeling) could result in extrinsic compression of the shunt. However, there is no evidence to support this scenario based on multiple observations of frame geometry seen during pre-clinical studies and during follow-up transesophageal echocardiography (TEE), CT, or fluoroscopic imaging in humans. In all cases, the observed shunt frame has not been observed to be extrinsically compressed or in any other way narrowed, deformed, or fractured.
Another possible mechanism is in situ thrombosis of the shunt. However, all patients were treated with monitored anticoagulation for the first three months, or indefinitely if there were other indications for chronic anticoagulation, which was most commonly required in patients with a history of atrial fibrillation. Subjects were also treated simultaneously with low-dose aspirin, which was continued indefinitely. Having experience with prosthetic cardiac valves as a predicate, valve thrombosis would have been expected to be seen earlier, typically within 30-45 days after implantation, especially in patients with a history of subtherapeutic anticoagulation therapy.
In the 38 patients implanted with the V-Wave valved hourglass-shaped shunt described above, no thrombi were detected on 121 consecutive post-implantation echocardiograms. These studies systematically looked for intracardiac or device thrombus by an independent Echocardiographic Core Laboratory at time points including one day after implantation, and at 1, 3, 6, and 12 months after implantation. None of the patients presented with stroke or other clinical manifestations of thromboembolic events. Of 9 patients with suspected shunt occlusion or incompetent valves, most were taking therapeutic doses of anticoagulants (warfarin or New Oral Anticoagulant agents) at the time the shunt anomaly was discovered. Another reason that thrombosis is unlikely is the observation of progressive vena contracta narrowing over a time course of 6 months or more. Thrombosis would be expected to result in sudden lumen loss, and not progress slowly over a period of months.
A third potential cause of occlusion is neoendocardial tissue overgrowth or pannus formation that narrows the lumen at the neck of the hourglass-shaped shunt. Applicants' earlier ovine studies suggest otherwise. Specifically, the shunt lumen surface at the neck of the hourglass contained only microscopic amounts of cellular material. On gross pathological examination, there was no visible loss of the lumen area in neck region. A human shunt specimen has been examined in an explanted heart from a patient that underwent cardiac transplantation 2.5 years after shunt implantation. The ePTFE surfaces of the shunt including the lumen at the neck contained no pannus formation or narrowing of any kind.
In another example, a left atrial pressure sensor implanted across the FO by transseptal catheterization and used for guiding the medical therapeutic dosing in symptomatic patients with severe heart failure was observed to experience pannus formation. In the original embodiment of the sensor, the sensing diaphragm, located at the distal end of the sensor module body, protruded into the left atrium by 1-mm beyond its three anchoring legs that rested on the left atrial side of the septum. In a later, improved geometry version, the legs were placed more proximal on the sensor module body so that sensing diaphragm protruded into the LA by an additional 1.5 mm.
In a comparative inter-species pathology study, neoendocardial tissue (pannus) formation was observed over the sensing diaphragm in 20 of 31 original sensors compared with only 3 of 40 specimens with the improved geometry sensor. Of the 20 original sensors with tissue coverage, 7 had demonstrable artifacts in the LA pressure waveform. In each case with artifacts, pannus formation over the sensing diaphragm had a thickness >0.3 mm. These data indicate that when tissue coverage exceeds this thickness, the tissue interferes with fluid pressure measurement. None of the improved sensors had waveform artifacts or tissue thickness >0.3 mm.
In addition to producing waveform artifacts, the time course of tissue encapsulation of the sensing diaphragm could be estimated by assessing LA pressure waveforms for baseline drift with or without the development of artifacts. It was hypothesized that as neoendocardial tissue grows over the sensing diaphragm, measured LA pressure increased due to a drifting baseline caused by tension applied from the tissue capsule covering the diaphragm through its contiguous connection with the atrial wall. This healing phenomenon may be initiated as early as several weeks' post implant in animals and starts around 3-4 months in humans. Using the timing of drift to indicate tissue coverage in humans, it was shown that in a group of 46 heart failure patients with the original sensor design geometry, about 25% developed the characteristic drift pattern associated with tissue coverage of the sensing diaphragm during the first year after implantation. Of 41 similar patients implanted with the improved geometry sensor, none developed drift.
Pannus formation on devices that traverse the interatrial septum has been observed to start at the portions of the device in contact with the septum in the region of local tissue injury. Tissue growth progresses contiguously, extending translationally along the external surfaces of the device that protrude into each atrial chamber. This pannus growth thins as a function of distance from the sites of cardiac contact until it becomes essentially a monolayer of neoendothelial cells. The process naturally stops after about 6-12 months in humans. Thereafter, the remaining tissue may remodel but active growth of pannus is completed. From these data, Applicants observed that tissue coverage typically grows a distance of about 3 mm from its starting place on the septal wall before stopping or becoming thin enough so as not to impede device function.
Thus, for pannus to cause narrowing of the lumen at the shunt neck, it would have to extend contiguously from the site of injury on the septum for some distance to reach the neck. Applicants have determined that translational tissue growth over a distance of 3 or more millimeters becomes much less likely.
Pannus formation affecting the valve leaflets is the most likely stand-alone mechanism that explains all of the untoward observations seen in human subjects implanted with V-Wave shunts, including progressive shunt narrowing, incompetence of the valve with bidirectional flow, and eventual loss of shunt flow with associated loss of clinical efficacy.
Tissue overgrowth affecting the valve leaflets bases and commissures was the predominant histopathological finding in the ovine pre-clinical study described above. Gross pathological examination of shunts implanted for 3 months showed pannus infiltration extending from the adjacent FO into the valve leaflet bases with thickening of the leaflet bodies in 5 out of 6 shunts. In 4 shunts, there was fusion of at least 2 of the 3 valve commissures where the leaflet edges were sutured to the shunt frame. Fusion of all 3 commissures was observed in 3 shunts. One case showed severe narrowing at the commissures with a luminal area of 4 mm2 or a 75% area stenosis in comparison to the normal 19.6 mm2 lumen at the device neck. The leaflets were described as semi-pliable or stiffened in 4 out of 6 shunts. In two of the devices, commissural fusion and leaflet thickening were so pronounced that complete leaflet coaptation could not likely occur during valve closure. In none of these cases has pannus formation been seen to narrow the shunt neck.
On examination of microscopic sections, pannus thickness tends to be greater on the side of the leaflets facing the atrial septum where the ePTFE/leaflet junction was infiltrated with pannus that was contiguous with the adjoining atrial tissue. Pannus extended from the atrial septum on and around the right atrial edge of the ePTFE skirt and into the base and commissures of the valve leaflets. At 3 months, the pericardial leaflets showed varying degrees of pannus coverage ranging from mild to marked. In general, pannus is thickest at the leaflet bases and commissures, and tapers toward the free edges. In 2 sheep, the pannus on the leaflets measured 2 to 3 times the original thickness of the leaflets.
The pannus was generally well healed or organized by 3 months. It was composed of collagen and proteoglycan matrix surrounding smooth muscle cells, fibroblasts and rare focal areas of inflammation with lymphocytes, macrophages, and occasional multinucleated (foreign body type) giant cells. The pannus tissue was mostly covered with neoendothelium consistent with near complete healing. No leaflet calcification or thrombi were observed.
Although animal models of cardiovascular devices are limited in their ability to represent human tissue healing responses, the major differences are characteristically limited to the temporal duration of the response. For example, in a comparative pathology study described in the publication, “Comparative Pathology of an Implantable Left Atrial Pressure Sensor” by Roberts, et al., of a percutaneously implantable titanium/nitinol-enclosed LA pressure sensor, implanted on the interatrial septum, it was found that sheep at 1.5 to 8 months and canines implanted for 1 to 25 months, closely approximated the pathological findings seen in humans implants of 3 to 56 months duration. Histology had a similar appearance in humans and animals, and confirmed that the tissue covering the device was composed of a neoendocardium lined with a neoendothelium. The appearance of the neoendocardial tissue covering the sensor described above was similar to what is observed with ASD closure devices.
This mechanism of pannus formation preferentially affecting the bioprosthetic valve material compared to the ePTFE encapsulated portions of the shunt was observed in the human explanted specimen referred to earlier. After 2.5 year of implantation in heart, the 3 pericardial leaflets were severely thickened, immobile, infiltrated at their bases and commissures with pannus resulting in valvular stenosis with a reduction in outflow area of 52% relative to the non-obstructed shunt neck. Although this shunt was patent, it would have been incompetent, allowing bidirectional flow, and would have shunted less than half of the flow expected for any given pressure gradient.
To further evaluate the tendency of this bioprosthetic valve to become infiltrated with pannus, valved and valveless designs of the V-Wave shunt were implanted by applicants in a non-diseased juvenile ovine (n=9) model. Specifically, this study was designed to highlight the resistance of a valveless, ePTFE encapsulated shunt (n=6) to pannus formation, narrowing and occlusion, relative to the legacy valved version previously used in humans (n=3), by creating a highly proliferative model expected in healthy juvenile sheep where the left-to-right interatrial pressure gradient was expected to be small. In the valveless design, the bioprosthetic valve material and its attaching polypropylene suture were removed and the ePTFE encapsulation was extended to cover the entire nitinol frame of the shunt except for the last 1.5 mm on the RA side where the shunt was coupled to its delivery system for deployment. The ePTFE used had an internodal distance of up to 30 microns. At 12 weeks the sheep where euthanized. The gross pathology findings showed that the 3 valved shunts were heavily infiltrated with pannus formation, extending from the septum into the regions containing the bioprosthetic leaflets. The leaflets were fused, immobile and highly stenotic leaving only a pinhole opening. The degree of pannus formation was much exaggeration versus prior experience in the ovine heart failure model. Thick pannus extended retrograde contiguously from the leaflet bases toward the hourglass neck of the shunts. The pannus growth from the original septal site of injury to the tips of the valve leaflets exceeded 3 mm in distance. Pannus appeared to grow through the valve commissures and through the suture holes attaching the porcine pericardial leaflets to the frame and the ePTFE skirt. Pannus formation was associated with mononuclear inflammatory cell infiltrates and multinucleated giant cells.
All 6 of the valveless, ePTFE encapsulated shunts were widely patent with only minimal pannus formation attaching the FO tissue to the external surface of the device. Applicants observed no pannus growing translationally more than 3 mm along the external surface of the ePTFE from the septum. No visible pannus reached from the septum all the way into the lumen portion of either the left atrial entry cone or right atrial exit cone of the device. The lumina at the necks of all of the shunts were widely patent on gross and microscopic examination. There was no evidence of pannus formation permeating through the ePTFE encapsulation into the shunt lumen.
From these combined observations, applicants have determined that length of translational pannus growth from the site of healing may be dependent on the type of biomaterial surface. In the case of the ePTFE encapsulated shunt, pannus formation severe enough to interfere with device function tends to translate a maximum of about 3 mm from the site of injury, whereas in the case of the bioprosthetic valve material tested, the amount of pannus formation and translational length of pannus tissue growth were exaggerated.
Also, from these data, it is reasonable to expect that the near complete shunt healing seen after 3 months in the juvenile ovine model will be predictive of the histopathological findings in humans at 9-12 months. Moreover, these gross and microscopic observations, with their anticipated species-to-species conservation of findings, leads to the conclusion that the healing response in sheep is likely indicative of the mechanism causing shunt closure, valvular incompetence, and progressive stenosis in humans. Thus, there exists a need for a more durable shunt configuration that maintains luminal patency for extended periods of time.
It further would be desirable to provide a shunt for redistributing atrial blood volumes and reducing interatrial pressure imbalances that reduces the risk of paradoxical embolism caused by emboli transfer from the right to left atria.
It also would be desirable to provide an interatrial shunt configuration that reduces the risk of pannus formation after a prolonged period of implantation, where the degree of pannus formation and tissue ingrowth is not strongly dependent on the manner or location in which the shunt is implanted in the FO.